arXiv subjects

Baichuan Mo

Publications and source records attributed to Baichuan Mo.

At least 19 recordsLinked to original sources

arXiv2026

A Dual-Guided Column-Generation-Based Matheuristic for Integrated Railroad Blocking

Integrated railroad blocking is a central planning problem in freight railway operations, which requires coordinated decisions on block design, blocking sequences, and block routing. We propose a dual-guided column-generation-based matheuristic (DG-CGM) that uses linear programming (LP) dual information to identify promising transportation paths and incorporates them into restricted mixed-integer programs to improve feasible solutions. Starting from an initial plan built by construction, DG-CGM alternates destroy and local repair with path generation, which uses dual prices from restricted LPs of a Dantzig-Wolfe reformulation to identify promising transportation paths. When local improvements continue to stagnate, we further apply joint repair to combine alternatives across demand groups. To address the coupling between logical and physical path pricing, we introduce a joint reduced cost that eliminates path-specific dual variables from pricing, enabling more efficient column identification through exact labeling algorithms. We evaluate the performance of DG-CGM on both generated instances and real-world cases. On 150 generated instances with 50-150 nodes, DG-CGM achieves the optimal cost on 89 of the 91 instances solved to proven optimality by Gurobi, with a geometric mean runtime equal to 5.2% of that for Gurobi and a mean relative cost reduction of 3.4%. On six full-scale scenarios derived from the 2026 INFORMS Railway Applications Section competition data, each with 47,193 nodes, DG-CGM obtains feasible solutions within two hours that serve 95.9%-99.6% of the requested freight volume, providing practical blocking plans for large-scale railway operations. The code, instances, and computational results are publicly available at https://github.com/Zhengzhong-You/IRBP.

math.OC↗

arXiv2026

Ozone: A Unified Platform for Transportation Research

Transportation research increasingly depends on heterogeneous data from roadside cameras, UAV imagery, LiDAR, and in-vehicle sensors. Differences in data representations, model interfaces, and evaluation protocols complicate cross-dataset benchmarking and reproducibility. We present Ozone, a unified transportation research platform organized around Hardware, Data, Model, and Evaluation layers, with a Prototype layer supporting deployment-oriented applications. Its data layer links vehicle trajectories, traffic conflicts, reconstructed crashes, and digital-twin maps through a shared trajectory schema. The source inventory spans 19 trajectory datasets recorded at 59 sites in eight countries, with completed converters for 12 sources and adaptation ongoing for seven. The converters align coordinate systems, object classes, oriented bounding boxes, and kinematic fields. A shared implementation of surrogate safety measures supports conflict extraction, while map and motion constraints link crash reports to reconstructed trajectories. CARLA digital-twin maps and SUMO traffic logic connect these records to simulation, model development, and evaluation. Validation on 500 vehicles from four datasets reports mean position, heading, and speed errors below 0.15 m, 2.5°, and 0.3 m/s. A workflow study reports an 85 % reduction in experiment setup time. Case studies demonstrate human-factor experiments and automated-driving safety testing, while the model interfaces support traffic modeling and world-model research. The source code and datasets are publicly available.

cs.DB↗

arXiv2026

Retrieval-Augmented Large Language Model Decision-Making for Autonomous Driving Guided by Chinese Philosophical Wisdom

Autonomous driving decision systems must balance safety, efficiency, and social norms in complex traffic interactions. Philosophical and ethical considerations have received limited attention in existing autonomous driving decision-making approaches based on numerical optimization, sequence prediction, and large language models (LLMs). We propose Chinese Philosophical Wisdom-Guided Driving (CPW-Drive), a closed-loop retrieval-augmented generation (RAG) framework that incorporates value guidance derived from Chinese philosophy into autonomous driving decision-making. Using Chinese Confucian thought as its knowledge source, CPW-Drive consolidates LLM-extracted keywords from relevant classical texts into driving-relevant value principles through manual screening and validation. It then contextualizes these principles through scenario-specific cases to form retrievable and reusable value guidance. We further propose Physics-aware Spatial Similarity Retrieval (PSSR), which compares vehicle layouts and velocity-extrapolated states to retrieve physically relevant historical cases. On Highway-env's multilane highway-driving task, CPW-Drive achieves success rates of 93.0%, 86.0%, and 72.0% across three traffic configurations. These results outperform the strongest baseline by 8.0, 22.5, and 25.0 percentage points, respectively. Across all configurations, CPW-Drive achieves the highest collision-free step count and maintains a low lane-change frequency. The results suggest that structured value guidance can improve simulated closed-loop safety and stability while introducing efficiency and latency trade-offs.

cs.AI↗

arXiv2026

Evolving in Thought Space: Training a Small Model at Test Time Unlocks Better Discoveries

Open-ended scientific discovery often requires repeatedly proposing and evaluating candidate solutions. LLM-based systems can support this process by generating and refining executable solutions from verifier feedback. Methods such as TTT-Discover use test-time training (TTT) to update the solution-generating LLM from verifier feedback, adapting its generation policy to improve subsequent proposals on the target problem. However, this becomes expensive when reliable execution requires a large model, since training must maintain gradients, optimizer states, and policy statistics while repeatedly generating long, structured outputs. It also complicates credit assignment: outcome-level verifier feedback must jointly evaluate the high-level strategy and its low-level implementation. In this work, we introduce Guidance-TTT, which separates these roles. A compact guidance model is trained at test time to propose high-level strategic changes, while a frozen execution model implements them as complete executable solutions. At each step, the system selects a promising previously discovered solution, proposes a change, executes and verifies it, and updates only the guidance model using an adaptive group-relative RL objective. This concentrates test-time learning on short strategic decisions while retaining the implementation capability of a substantially stronger model without adapting it. Without web access, Guidance-TTT produces strong solutions across four distinct domains: combinatorial optimization (Polyomino Packing), heuristic programming (AHC058), machine learning (Lasso), and GPU kernel optimization (TriMul). Across these tasks, it outperforms the best solutions reported in prior work while remaining competitive with state-of-the-art results on public online leaderboards. Code is available at https://github.com/Human-Agent-Society/reef/tree/guidance-ttt-support.

cs.AI↗

arXiv2026

PrecPack: An Efficient Open-Source Exact Solver for Bin Packing with Generalized Precedence Constraints

Efficient resource use in packing and assembly-line applications requires decisions that jointly account for capacity and precedence constraints. The strongly NP-hard bin packing problem with generalized precedence constraints (BPP-GP) models such decisions by minimizing the number of ordered, capacitated bins required to pack weighted items, even when precedence requirements span multiple bins. Existing exact algorithms primarily focus on classical special cases, whereas general BPP-GP has been addressed only via compact integer models and heuristics, with no efficient open-source exact solver. We present PrecPack, a unified exact solver that extends branch-bound-and-remember (BBR) to arbitrary nonnegative precedence weights and naturally specializes to the classical cases. Generalized states capture restrictions that remain active across future bins, which are addressed through branching, dominance, and conflict-aware lower bounds. Root column generation uses fixed-point arithmetic to compute numerically valid dual bounds for pruning or to prove optimality. To support reuse and verification, we provide common programming and command-line interfaces, independent assignment checking, explicit termination statuses, and reproducible batch execution; the core procedures require no commercial software. In same-machine, single-threaded comparisons on classic assembly-line benchmarks, more instances are proven optimal, and average computing times are substantially reduced relative to leading source-available BBR implementations. Further comparisons with published benchmark results for bin packing with precedence constraints and BPP-GP also show that more instances were proved optimal and that reported average gaps were smaller on most benchmark sets. PrecPack is released under the MIT License at https://github.com/Sunkanghong-Wang/PrecPack.

cs.DS↗

arXiv2026

When Do Corrective Features Help? An Agent for Corrective Feature Discovery on Black-Box Forecasters

Frozen pretrained forecasters often fail in structured, recurring ways that are costly to repair through fine-tuning. We study corrective feature discovery: mining interpretable features of a frozen forecaster's residual to drive a lightweight post-hoc corrector. Prior automated feature engineering models the data-generating process; corrective features instead model the model-failure process. We present CRAFTER (Corrective Residual Agent with Feature-based Temporal Exploration and Reasoning), which keeps the backbone frozen and mines its residual with two complementary generators: a compositional search over the raw input channels, and a large language model (LLM) that proposes named feature combinations, binary flags, and short executable code. A single validation-grounded gate accepts or rejects every candidate regardless of its origin, and a validation-selected corrector applies the accepted features or leaves the forecast unchanged. This source-agnostic pipeline also allows prior feature-engineering systems to be evaluated under identical conditions, making CRAFTER an instrument for attributing forecast improvements to the feature source alone. Across six public datasets and six frozen backbones, CRAFTER surpasses every dedicated feature-engineering system at every feature budget, roughly doubling the improvement achieved by the corrector alone and reducing the error of the weakest backbones by up to 27%. These gains are robust across different LLM backends and persist even when applied on top of fine-tuned backbones.

cs.LG↗

arXiv2026

TorchDCM: A Unified PyTorch-Native Package for Discrete Choice Modeling

Estimating large and simulation-intensive discrete choice models (DCMs) requires repeated evaluation of utilities, probabilities, derivatives, and simulated likelihoods over many observations, alternatives, and draws. Existing DCM software provides mature econometric workflows, while recent GPU-oriented tools accelerate selected models, leaving a gap between econometric coverage and scalable differentiable computation. We introduce TorchDCM, an open Python package for discrete choice modeling that compiles choice data and model specifications into a unified PyTorch-native likelihood engine for estimation, inference, prediction, and structured reporting on CPU or CUDA devices. The package covers the principal econometric functionality available across Biogeme and Apollo, including multinomial, nested, mixed, ordered, latent-variable, and panel likelihoods. It also supports ragged choice sets, constrained parameters, covariance estimation, willingness-to-pay analysis, elasticities, and extensible likelihood components. We evaluate TorchDCM against seven other estimation packages in aligned synthetic and real-data full-estimation experiments. TorchDCM completes all 45 synthetic cases, runs fastest in every comparable synthetic case, and satisfies the prespecified final-log-likelihood tolerance in every comparison with at least two comparable solutions. More precisely, it reduces median runtime by 89.1%-99.7% relative to Biogeme and Apollo across model-data settings. CUDA provides an additional 12.0-71.0x speedup over single-core TorchDCM. These results establish a scalable and reproducible foundation for econometric estimation and differentiable choice-model development. The open-source package and executed examples are available at https://github.com/mbc96325/torchdcm.

stat.ML↗

arXiv2026

Bulk Service Queueing for Transit Resilience under Short Random Service Suspensions

Short service suspensions are common in public transit systems, but their operational impacts remain difficult to quantify. We develop an analytical framework for measuring the resilience of a transit line under short random service suspensions. Vehicle movement is represented by a two state process in which vehicles either travel normally or stop during a suspension, and the induced stochastic headways enter a bulk service queueing model with finite vehicle capacity and passenger carryover. The model yields two classes of resilience indicators. Stability conditions determine whether station queues remain bounded, while closed form expressions characterize the mean and variance of station level queue length and waiting time. We construct an independent renewal approximation for headways whose common marginal distribution is obtained by taking the positive part of a raw headway formed from the incident adjusted scheduled headway and the difference between two independent compound Poisson exponential variables. The renewal approximation preserves the marginal effects of short suspensions while omitting serial dependence and delay propagation across multiple vehicles. Combining the resulting passenger arrival distribution with a Markov representation of passenger loads across stations allows the resilience indicators to be computed sequentially along the route. Numerical experiments show that short suspensions disproportionately affect congested stations and that changes in incident duration and scheduled headway can dominate comparable changes in vehicle capacity. A recursive first in, first out simulation assesses the analytical approximations and clarifies the role of headway variability.

math.PR↗

arXiv2026

Learned Pairwise Deep Dual-Optimal Inequalities for Stabilizing Column Generation

Column generation (CG) is central to many large-scale optimization algorithms, including branch-price-and-cut methods for vehicle routing problems, but unstable dual solutions can substantially slow its convergence. Existing deep dual-optimal inequalities can reduce this instability by restricting the dual space. Their construction, however, typically relies on problem-specific exchange arguments that are difficult to establish for routing problems with capacity limits, time windows, and other resource constraints. We introduce learned pairwise deep dual-optimal inequalities (L-PDDOIs), a learning framework that predicts pairwise orderings between dual variables and incorporates their primal counterparts directly into the master problem. To construct training labels, the framework samples optimal dual solutions and selects pairwise order relations that hold simultaneously on a sufficiently large common subset of the samples. A classifier then assigns a score to each candidate relation. Because conflicts and redundancies among the predicted relations can impair performance, graph-based postprocessing filters and compresses the candidate set before deployment. We further introduce a recovery procedure that selectively relaxes learned inequalities and provides a certificate when the baseline CG bound has been restored. On the main test sets for the capacitated vehicle routing problem and the vehicle routing problem with time windows, direct deployment of L-PDDOIs reduces the geometric mean root CG time by 89.7% and 93.9%, respectively, while incurring mean bound losses of only 1.3% and 0.5%. The recovery procedure retains corresponding time reductions of 54.8% and 83.1%, respectively, while guaranteeing no loss in the CG bound.

math.OC↗

arXiv2026

Large Language Models for Travel Behavior Prediction

Travel behavior prediction is a core problem in transportation demand management and is traditionally addressed using numerical models calibrated on observed data. With recent advances in large language models (LLMs), new opportunities have emerged to model human decision-making through natural language reasoning. This study explores the use of LLMs for travel behavior prediction through two complementary frameworks. The first framework employs a zero-shot prompting strategy, where the prediction task, traveler attributes, and relevant domain knowledge are described in text, enabling the LLM to directly generate predictions without task-specific training data. The second framework uses LLM-generated text embeddings as high-level representations of travel scenarios, which are then combined with conventional supervised learning models to support prediction in small-sample settings. Empirical results show that both approaches achieve performance comparable to, and in some cases competitive with, classical models such as multinomial logit, random forest, and neural networks. These findings suggest that LLMs offer a flexible and data-efficient alternative for travel behavior prediction.

cs.LG↗

arXiv2026

Online Linear Programming with Replenishment

We study an online linear programming (OLP) model in which inventory is not provided upfront but instead arrives gradually through an exogenous stochastic replenishment process. This replenishment-based formulation captures operational settings, such as e-commerce fulfillment, perishable supply chains, and renewable-powered systems, where resources are accumulated gradually and initial inventories are small or zero. The introduction of dispersed, uncertain replenishment fundamentally alters the structure of classical OLPs, creating persistent stockout risk and eliminating advance knowledge of the total budget. We develop new algorithms and regret analyses for three major distributional regimes studied in the OLP literature: bounded distributions, finite-support distributions, and continuous-support distributions with a non-degeneracy condition. For bounded distributions, we design an algorithm that achieves $\widetilde{\mathcal{O}}(\sqrt{T})$ regret. For finite-support distributions with a non-degenerate induced LP, we obtain $\mathcal{O}(\log T)$ regret, and we establish an $Ω(\sqrt{T})$ lower bound for degenerate instances, demonstrating a sharp separation from the classical setting where $\mathcal{O}(1)$ regret is achievable. For continuous-support, non-degenerate distributions, we develop a two-stage accumulate-then-convert algorithm that achieves $\mathcal{O}(\log^2 T)$ regret, comparable to the $\mathcal{O}(\log T)$ regret in classical OLPs. Together, these results provide a near-complete characterization of the optimal regret achievable in OLP with replenishment. Finally, we empirically evaluate our algorithms and demonstrate their advantages over natural adaptations of classical OLP methods in the replenishment setting.

math.OC↗

arXiv2025

Predicting Drivers' Route Trajectories in Last-Mile Delivery Using A Pair-wise Attention-based Pointer Neural Network

In last-mile delivery, drivers frequently deviate from planned delivery routes because of their tacit knowledge of the road and curbside infrastructure, customer availability, and other characteristics of the respective service areas. Hence, the actual stop sequences chosen by an experienced human driver may be potentially preferable to the theoretical shortest-distance routing under real-life operational conditions. Thus, being able to predict the actual stop sequence that a human driver would follow can help to improve route planning in last-mile delivery. This paper proposes a pair-wise attention-based pointer neural network for this prediction task using drivers' historical delivery trajectory data. In addition to the commonly used encoder-decoder architecture for sequence-to-sequence prediction, we propose a new attention mechanism based on an alternative specific neural network to capture the local pair-wise information for each pair of stops. To further capture the global efficiency of the route, we propose a new iterative sequence generation algorithm that is used after model training to identify the first stop of a route that yields the lowest operational cost. Results from an extensive case study on real operational data from Amazon's last-mile delivery operations in the US show that our proposed method can significantly outperform traditional optimization-based approaches and other machine learning methods (such as the Long Short-Term Memory encoder-decoder and the original pointer network) in finding stop sequences that are closer to high-quality routes executed by experienced drivers in the field. Compared to benchmark models, the proposed model can increase the average prediction accuracy of the first four stops from around 0.229 to 0.312, and reduce the disparity between the predicted route and the actual route by around 15%.

cs.LG↗

arXiv2025

Individual Path Recommendation Under Public Transit Service Disruptions Considering Behavior Uncertainty

Public transit passengers need guidance during service disruptions. This study proposes an individual-based path (IPR) recommendation model. The model decides which paths to recommend for each passenger with the objective of minimizing system travel time and respecting passengers' path choice preferences. We assume the recommendations could affect passengers' path choice probabilities, but their actual choices are uncertain. This behavior uncertainty makes the problem a stochastic optimization with decision-dependent distributions. We propose a single-point approximation method to eliminate the expectation operator by introducing two new concepts: epsilon-feasibility and Gamma-concentration, which control the mean and variance of path flows in the optimization problem. The approximation yields a tractable single-stage mixed integer linear formulation, which can be solved efficiently with Benders decomposition. The approximation gap is approved to be bounded from the above. Additional theoretical analysis shows that epsilon-feasibility and Gamma-concentration are strongly connected to expectation and chance constraints in a typical stochastic optimization formulation, respectively. The model is implemented in a real-world case study using data from an urban rail disruption in the Chicago Transit Authority system, and a synthetic case study with varied network sizes and incident locations. In the real-world case study, results show that the proposed IPR model reduces the average travel times in the system by 6.6% compared to the status quo and by 4.2% compared to a capacity-based benchmark model. In the synthetic case study, the proposed model shows 15.0% to 1.8% lower system travel time compared to the capacity-based benchmark, depending on the network sizes and demand situations.

math.OC↗

arXiv2025

Robust Binary and Multinomial Logit Models for Classification with Data Uncertainties

Binary logit (BNL) and multinomial logit (MNL) models are the two most widely used discrete choice models for travel behavior modeling and prediction. However, in many scenarios, the collected data for those models are subject to measurement errors. Previous studies on measurement errors mostly focus on "better estimating model parameters" with training data. In this study, we focus on using BNL and MNL for classification problems, that is, to ``better predict the behavior of new samples'' when measurement errors occur in testing data. To this end, we propose a robust BNL and MNL framework that is able to account for data uncertainties in both features and labels. The models are based on robust optimization theory that minimizes the worst-case loss over a set of uncertainty data scenarios. Specifically, for feature uncertainties, we assume that the l_p-norm of the measurement errors in features is smaller than a pre-established threshold. We model label uncertainties by limiting the number of mislabeled choices to at most Gamma. Based on these assumptions, we derive a tractable robust counterpart. The derived robust-feature BNL and the robust-label MNL models are exact. However, the formulation for the robust-feature MNL model is an approximation of the exact robust optimization problem. An upper bound of the approximation gap is provided. We prove that the robust estimators are inconsistent but with a higher trace of the Fisher information matrix. They are preferred when out-of-sample data has errors due to the shrunk scale of the estimated parameters. The proposed models are validated in a binary choice data set and a multinomial choice data set, respectively. Results show that the robust models (both features and labels) can outperform the conventional BNL and MNL models in prediction accuracy and log-likelihood.

math.OC↗

arXiv2025

TimeMixer++: A General Time Series Pattern Machine for Universal Predictive Analysis

Time series analysis plays a critical role in numerous applications, supporting tasks such as forecasting, classification, anomaly detection, and imputation. In this work, we present the time series pattern machine (TSPM), a model designed to excel in a broad range of time series tasks through powerful representation and pattern extraction capabilities. Traditional time series models often struggle to capture universal patterns, limiting their effectiveness across diverse tasks. To address this, we define multiple scales in the time domain and various resolutions in the frequency domain, employing various mixing strategies to extract intricate, task-adaptive time series patterns. Specifically, we introduce a general-purpose TSPM that processes multi-scale time series using (1) multi-resolution time imaging (MRTI), (2) time image decomposition (TID), (3) multi-scale mixing (MCM), and (4) multi-resolution mixing (MRM) to extract comprehensive temporal patterns. MRTI transforms multi-scale time series into multi-resolution time images, capturing patterns across both temporal and frequency domains. TID leverages dual-axis attention to extract seasonal and trend patterns, while MCM hierarchically aggregates these patterns across scales. MRM adaptively integrates all representations across resolutions. This method achieves state-of-the-art performance across 8 time series analytical tasks, consistently surpassing both general-purpose and task-specific models. Our work marks a promising step toward the next generation of TSPMs, paving the way for further advancements in time series analysis.

cs.LG↗

arXiv2025

Robust Path Recommendations During Public Transit Disruptions Under Demand Uncertainty

When there are significant service disruptions in public transit systems, passengers usually need guidance to find alternative paths. This paper proposes a path recommendation model to mitigate congestion during public transit disruptions. Passengers with different origins, destinations, and departure times are recommended with different paths such that the system travel time is minimized. We model the path recommendation problem as an optimal flow problem with uncertain demand information. To tackle the lack of analytical formulation of travel times due to capacity constraints, we propose a simulation-based first-order approximation to transform the original problem into a linear program. Uncertainties in demand are modeled using robust optimization to protect the path recommendation strategies against inaccurate estimates. A real-world rail disruption scenario in the Chicago Transit Authority (CTA) system is used as a case study. Results show that even without considering uncertainty, the nominal model can reduce the system travel time by 9.1% (compared to the status quo), and outperforms the benchmark capacity-based path recommendation. The average travel time of passengers in the incident line (i.e., passengers receiving recommendations) is reduced more (-20.6% compared to the status quo). After incorporating the demand uncertainty, the robust model can further reduce system travel times. The best robust model can decrease the average travel time of incident-line passengers by 2.91% compared to the nominal model. The improvement of robust models is more prominent when the actual demand pattern is close to the worst-case demand.

math.OC↗

arXiv2025

Comparing hundreds of machine learning classifiers and discrete choice models in predicting travel behavior: an empirical benchmark

Numerous studies have compared machine learning (ML) and discrete choice models (DCMs) in predicting travel demand. However, these studies often lack generalizability as they compare models deterministically without considering contextual variations. To address this limitation, our study develops an empirical benchmark by designing a tournament model, thus efficiently summarizing a large number of experiments, quantifying the randomness in model comparisons, and using formal statistical tests to differentiate between the model and contextual effects. This benchmark study compares two large-scale data sources: a database compiled from literature review summarizing 136 experiments from 35 studies, and our own experiment data, encompassing a total of 6,970 experiments from 105 models and 12 model families. This benchmark study yields two key findings. Firstly, many ML models, particularly the ensemble methods and deep learning, statistically outperform the DCM family (i.e., multinomial, nested, and mixed logit models). However, this study also highlights the crucial role of the contextual factors (i.e., data sources, inputs and choice categories), which can explain models' predictive performance more effectively than the differences in model types alone. Model performance varies significantly with data sources, improving with larger sample sizes and lower dimensional alternative sets. After controlling all the model and contextual factors, significant randomness still remains, implying inherent uncertainty in such model comparisons. Overall, we suggest that future researchers shift more focus from context-specific model comparisons towards examining model transferability across contexts and characterizing the inherent uncertainty in ML, thus creating more robust and generalizable next-generation travel demand models.

cs.LG↗

arXiv2023

Amazon Last-Mile Delivery Trajectory Prediction Using Hierarchical TSP with Customized Cost Matrix

In response to the Amazon Last-Mile Routing Challenge, Team Permission Denied proposes a hierarchical Travelling Salesman Problem (TSP) optimization with a customized cost matrix. The higher level TSP solves for the zone sequence while the lower level TSP solves the intra-zonal stop sequence. The cost matrix is modified to account for routing patterns beyond the shortest travel time. Lastly, some post-processing is done to edit the sequence to match commonly observed routing patterns, such as when travel times are similar, drivers usually start with stops with more packages than those with fewer packages. The model is tested on 1223 routes that are randomly selected out of the training set and the score is 0.0381. On the 13 routes in the given model apply set, the score was 0.0375.

math.OC↗